From Gambits to Assurances: Game-Theoretic Integration of Safety and Learning for Interactive Robotics

Autonomous robots are becoming more versatile and widespread in our daily lives. From autonomous vehicles to companion robots for senior care, these human-centric systems must demonstrate a high degree of reliability in order to build trust and, ultimately, deliver social value. How safe is safe enough for robots to be wholeheartedly trusted by society? Is it sufficient if an autonomous vehicle can avoid hitting a fallen cyclist 99.9% of the time? What if this rate can only be achieved by the vehicle always stopping and waiting for the human to move out of the way? I argue that, for trustworthy deployment of robots in human-populated space, we need to complement standard statistical methods with clear-cut robust safety assurances under a vetted set of operation conditions as well established as those of bridges, power plants, and elevators. We need runtime learning to minimize the robot's performance loss during safety-enforcing maneuvers by reducing its inherent uncertainty induced by its human peers, for example, their intent (does a human driver want to merge, cut behind, or stay in the lane?) or response (if the robot comes closer, how will the human react?). We need to close the loop between the robot's learning and decision-making so that it can optimize efficiency by anticipating how its ongoing interaction with the human may affect the evolving uncertainty, and ultimately, its long-term performance. My vision is to enable interactive robotic systems that can be built, deployed, and verified with safety assurances under minimal performance loss. Towards this goal I have developed new algorithms and theorems centered around dynamic game theory, integrating insights from robust optimal control, deep reinforcement learning, generative AI, and numerical optimization. The core of my program is to plan robot motion in the joint space of both physical and information states, actively ensuring safety as robots navigate uncertain, changing environments and interact with humans. A consistent principle throughout my research is to ensure that my methods can be validated with hardware tests and that they are reproducible by independent experts. The key contributions of my work include: 1. Trustworthy human-robot interaction: planning safe and efficient trajectories by closing the computation loop between interaction and runtime learning that actively reduces the robot's uncertainty about the human [1-7]. 2. Verifiable neural safety analysis for complex robotic systems: learning robust neural controllers for robots with high-dimensional dynamics; guaranteeing their training-time convergence and deployment-time safety [8-11]. 3. Scalable game-theoretic planning under uncertainty: collaborating with neuroscientists and operations research experts to develop new game-theoretic methods for complex and uncertain human-robot systems [12-16]. Impact. Together, these contributions lay the foundation for next-generation interactive robotic systems deployed with verifiable assurances and real-time adaptability in uncertain, unstructured, and human-populated environments. My work is nominated for the Roberto Tempo Best Paper Award at the IEEE Conference on Decision and Control (CDC), and has attracted global attention from industry leaders like Toyota Research Institute and Honda Research Institute who build on my research. In recognition of my contributions, I have been named as a Human-Robot Interaction Pioneer (rising young researcher) by IEEE and ACM, and appointed as an Associate Editor of IEEE Robotics and Automation Letters (RA-L), a rare honor for a Ph.D. student. Agenda. As a professor, I intend to lead a research agenda at the intersection of robotics and artificial intelligence, with a clear focus on enhancing public trust in human-centered robotics through transparent safety assurances and cutting-edge capabilities. I will equip my lab with testbeds that I have worked extensively with, including autonomous vehicles [1-4, 7, 12-15, 17], quadrotors [12, 18, 19], and legged robots [8, 11], to carry out both theoretical and experimental work that advances the fields. I will leverage my external collaborators across the industry (Toyota, Honda) and government (NSF, ONR, DARPA) to ensure that my research drives the success of academia and society at large. The long-term goal of my lab is to develop comprehensive safety principles that contribute to shaping regulatory standards and enhancing public trust in human-centered autonomy.